Human milk is the ideal food for newborns, especially if they are premature or ill,1–4 but its properties may alter during its manipulation and storage.

Freezing human milk at −20°C guarantees its microbiological quality, but its effects on nutritional quality have not been researched thoroughly. Some studies have observed a decrease in the bactericidal and antioxidant activities of breast milk5–7 and a reduction in the energy and fat content over the freezing time.8

At times, the nutritional content of breastmilk does not meet the requirements of preterm and low-birth-weight infants,9 so it is important that we learn its nutritional value. Lipids constitute the main source of energy, but they are the component that shows the highest degree of variation.10 The creamatocrit method is a simple and inexpensive means to calculate the fat and energy content of breast milk. It was described by Lucas et al.11 in 1978. Several authors have observed a good correlation between the energy content of milk and the creamatocrit.12–14

The purpose of our study was to determine, by means of the creamatocrit method, the variations in the fat content of raw and pasteurised human milk throughout 3 months of freezing at −20°C.

We performed an experimental study. We obtained the samples from women who donate to the Milk Bank, and gave informed consent. The samples were expressed manually or with a breast pump, and stored in sterilised glass containers.

The samples of raw (unpasteurised) milk were refrigerated at 4–5°C following extraction for a maximum of 24h. In those first 24h the samples were homogenised by rocking them in an arc-like fashion 10 times and then were divided in 7 aliquots. We analysed the first aliquot (time 0) and the 6 remaining aliquots were frozen at −20°C.

The samples of pasteurised milk were homogenised immediately following Holder pasteurisation by rocking them in an arc-like fashion 10 times, after which they were divided into 7 aliquots. We analysed the first aliquot after pasteurisation, and the 6 remaining aliquots were frozen at −20°C.

The raw and pasteurised milk aliquots were stored in polypropylene tubes and were labelled based on freezing time (7, 14, 21, 30, 60, or 90 days).

To analyse the aliquots, they were thawed in a 40°C water bath until a frozen pellet remained in the middle, and moved to a refrigerator to complete the thawing at 4–5°C.

We collected 44 samples of raw donated breastmilk. Table 1 shows the number and percentage of raw milk aliquots on which 1, 2, and 3 creamatocrits were performed.

The mean±standard deviation of the fat content at the initial time point was 3.19±1.44g/dl, and for the energy content it was 640.05±140.60kcal/l. We observed a significant decrease in the fat and energy content measurements after freezing at −20°C at all times of analysis (7, 14, 21, 30, 60, and 90 days) (Table 2).

We collected 36 samples of pasteurised donated milk. Table 3 shows the aliquots on which 1, 2, and 3 creamatocrits were performed.

The mean fat content at the beginning of the study was 2.59±1.17g/dl, and the mean energy content was 581.56±114.04kcal/l. We observed a significant reduction in the mean fat and energy content after freezing at −20°C during the first month of the study (7, 14, 21, and 30 days) (Table 4).

Our study showed a reduction in the fat and energy content measured by the creamatocrit method in raw and pasteurised breastmilk following freezing at −20°C. This reduction was significant for the first three months of the study in raw milk, and only for the first month in pasteurised milk. Until now, no previous study has assessed the changes in crematocrit values in either raw or pasteurised milk during 3 months of freezing.

Several studies describe an increase in milk lipase activity during freezing at −20°C.15–17 Lipolysis hydrolises triglycerides, increasing the level of free fatty acids in milk and reducing the size of the cream layer. This would account for the findings pertaining to the reduction in fat and energy content measured by the creamatocrit method in raw milk from the initial time point to 3 months of freezing. In our study, we also saw that the reduction in the measured fat content of pasteurised milk only occurred during the first month. It has been reported that pasteurisation reduces the levels of lipase activity,18 and a lower rate of triglyceride hydrolysis could explain why the reduction in the cream layer is smaller in pasteurised milk than in raw milk.

Therefore, while it is likely that the total fat content does not change during freezing, changes in the fat globule affect the creamatocrit value. Lucas et al.11 previously described that the breakdown of fat globules could produce a false decrease in the creamatocrit, although their study did not find significant changes in the creamatocrit after freezing for 2 months. In contrast, other authors14 described a decrease in the creamatocrit after 7 days of freezing at −20°C; in this study, the correlation between the creamatocrit and the lipid and energy content remained strong, but they expressed the need to develop a new formula to calculate the energy content of frozen milk more accurately.

The intermediate values that we observed in our study, both in raw and pasteurised milk, cannot be readily explained, and evince the limitations of the creamatocrit to measure the fat content of processed milk.

The main limitation of this study is that, due to technical problems, we could not make three creamatocrit measurements for each sample, as indicated by the Lucas method,11 although other studies have used duplicate creamatocrit measurements with good results.19,20

For all of the above, we believe that the creamatocrit, while being a method that is easily available in neonatal units, is not the optimal way to measure fat content in previously processed milk. One alternative would be to modify Lucas’ formula11 for its use in frozen or pasteurised milk, or other more valid measurement methods can be used.

The authors have no conflicts of interest to declare.